The Key Factors Controlling the Mechanisms of Consecutive Failures in the Collolar Open Pit in the Afsin-Elbistan Lignite (AEL) Basin in Southeast Türkiye

doi:10.21203/rs.3.rs-4359801/v1

Download PDF

Research Article

The Key Factors Controlling the Mechanisms of Consecutive Failures in the Collolar Open Pit in the Afsin-Elbistan Lignite (AEL) Basin in Southeast Türkiye

https://doi.org/10.21203/rs.3.rs-4359801/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Mass movements pose a great threat to the safe operation of open pit mines and seriously affect the mining economy, human health and environmental safety. In cases of insufficient dewatering or excessive groundwater withdrawal, different types of stability problems may arise especially in weak and saturated geological settings. In engineering practice, back analyses by finite element method (FEM) is a valuable tool for the evaluation of such cases and revealing their causes. In this study, mechanisms of consecutive failures occurred in an open pit slopes in the Afşin Elbistan Lignite (AEL) basin, southeast of Türkiye, were assessed based on intensive data obtained from the field and laboratory investigations. As a result of back analyses carried out through 2D FEM analyses, it was concluded that the high plasticity clay layers within the geological formation and the hydraulic interaction of two aquifers with different characteristics at the edge of the basin are two key factors controlling mining activities in the AEL basin.

Afşin Elbistan Lignite basin

2D FEM analysis

Landslide

Open pit

Sinkhole

In mining practice, numerous difficult-to-solve problems in difficult geological conditions have been encountered around the world. In such cases, even though the most sophisticated and geotechnically optimal options have been tried, successful results may not have been achieved and some catastrophic failures may not have been prevented. In general, the contradiction of geotechnical applications requiring high costs and other engineering constraints leads to the fact that the problem cannot be overcome.

Numerous instability problems in open pit mines, especially in the coalfield are evident. Practically most of these mines must deal with groundwater problems either before, during or after mining. The presence of groundwater in a mine site has adverse effects on mine production, slope stability, safety, pollution control and therefore mining cost. However, ground movements such as landslides and sinkholes that have occurred in recent years point to open-pit mining problems that could not be mitigated in soft, weak and saturated geological environments (Wolkersdorfer et al., 1999, Tzampoglou and Loupasakis, 2023, Mahmutoğlu et al., 2017, Mahmutoğlu et al., 2019, Lashgari and Öztürk, 2022, Tutluoğlu et al., 2011).

Zevgolis et al. (2019) have well documented slope failure events and other instability concerns associated with surface lignite mines in Greece and pointed to the importance of stratigraphic sequencing in lignite basins. Leonarda (2004) and Kavvadas et al. (2018) have pointed to slope instabilities usually governed by sliding along a sub-horizontal, unfavourably sloping towards the excavation, interface between lignite and an underlying stiff, high plasticity clay or marl layer in the same coalfields. Özbay and Cabalar (2015) have discussed the results of back analysis of slope failures in February 6 and February 10 of the Collolar open pit in 2011 and concluded that the first landslide at the permanent slopes occurred possibly due to the high-water table levels which was continuously fed by the Hurman River on the west. Akçar et al. (2019) have explained the development and propagation of the second landslide on February 10, 2011 (Landslide II in this study) as a rearward propagation failure of the thick and stiff Quaternary sequence over liquefied layer or layers of “seekreide” or “seekreide bearing clays” in the Collolar coalfield.

Gutierrez et al. (2008) have listed main changes in karst environments and mentioned that negative effects of natural processes and human activities may trigger or accelerate the development of sinkholes. Newton (1984) has estimated that more than 4000 of sinkholes, subsidence area or related features have formed since 1900 in Alabama, and has explained the mechanism in carbonate terranes beneath unconsolidated rocks. Thrap (1999) has examined cover collapse sinkholes which occur by failure of the walls of a solid above cavernous bedrock and explained mechanics of upward propagation of cover-collapse sinkholes for steady state and transient pore pressure conditions. Sevil and Gutierrez (2023) and Yechieli et al. (2016) suggested a similar mechanism for sinkhole formation along the Dead Sea coast. They have pointed to circulation of water from lower to higher sub-aquifers along cracks and faults.

The AEL basin in Southeast Türkiye is one of the rarest coalfields in the country in terms of reserve and low stripping ratio. A significant proportion of the country's lignite reserve is in this basin (Gürsoy et al., 1981 and 2011, Yörükoğlu, 1991). Karstified bedrock, overlined by the young and weak Tertiary aged terrestrial and lacustrine sediments with coal layer, is a confined aquifer with high efficiency controlling groundwater balance at the regional scale in this basin (Besbelli et al., 2009, Gökmenoğlu and Arslan, 2013, CU MMF, 2014, Karagüzel et al., 2018, Mahmutoğlu et al., 2019).

The young lacustrine deposit with coal is completely saturated, and a thick impermeable clay at the bottom of this sequence is a barrier to direct hydraulic interaction with the confined karst aquifer at the base in Collolar Sector. Therefore, excessive groundwater withdrawal from one of these aquifers can only cause rapid and direct interaction throughout their contacts at the edge of the basin. In this study, mechanisms of consecutive failures occurred in and around open pit slopes in the AEL basin were assessed based on intensive data obtained from the field and laboratory investigations. In first, the instability problems that have arisen from the start of mining activities in the Collolar open pit are introduced chronologically. Then, our evaluations based on our own field studies, existing and new research data from drilling, testing and monitoring surveys are given for the characterizations of the geological environment. Finally, site specific conditions, results of 2D FEM analyses of each consecutive instability, and the causes and mechanisms of failures are discussed.

The AEL basin has an area of approximately 900 km² within the Afşin and Elbistan districts of Kahramanmaraş province and is located on average 1200 mASL. The long axis of the AEL basin is in northwest-southeast direction and the surface drainage area is 2000 km². On the borders of the basin there are mountainous terrains that reach over 2500 m. Meteorological water balance assessments for the basin showed that the average annual precipitation is 439 mm, the actual evaporation value is 356 mm, 82.6 mm the total value of surface and underground flow is estimated and the amount of water percolating underground was 24.8 mm/year (Besbelli et al., 2009). The surface waters of the basin are drained by the Hurman River and its tributaries on the west of the open pit (Fig. 1).

In 1966 the first coal exploration in the AEL was carried out by the German Otto Gold and General Directorate of Mineral Research and Exploration of Türkiye. The reserve of lignite, which is estimated to be spread over a total area of 120 km², was determined as 3.4 billion tons (Gürsoy et al., 1981). Schloemer (1972) pointed to groundwater problems of the Neogene formations on top of coal layer in the Collolar Sector and stated the coal itself is impermeable.

In 2005, the permanent western slope of the Kislakoy open pit, east of Collolar Sector, which started to be operated in the 1980s, was disturbed, and then in 2006, 4 landslides covering the entire west and north slopes of open pit was occurred. Akbulut et al. (2013) stated firstly that the black clay bands with high plasticity present in the lignite horizon have the most crucial part in controlling the slopes stability. Tutluoğlu et al (2011) have also pointed to black clay as a critical geological layer for stope stability in Collolar Open Pit.

The surface mine operation plan in the Collolar Sector was prepared in 2007 and overburden stripping works were started in 2008. A drainage program was foreseen by Montan Bildungs und Entwicklungsgesellschaft mbH (MBEG, 2008) considering the hydrogeological characteristics of the geological units around the open pit. To ensure the stability of the open pit walls it was envisaged that the unconfined aquifer would be dewatered down to the top level of coal horizon before mining operations. The excavation works continued down to 100 m depth and coal level was reached in a short time, after which the first surface cracks showing instabilities were detected on SW slopes in 2009, then the first landslide occurred in the same pit wall in January 2010. Considering the negative interaction of the open pit and the Hurman River, the river bed was slightly moved behind the southwest slope after the landslides of February 2011 (Fig. 1).

3.1. The First Tension Cracks

The mining activities started in 2008 in the Collolar Sector focusing to reach immediately the coal needed by the existing thermal power plant. For this reason, the stripping works on coal were carried out very quickly, and the coal level was reached in June 2009. After excavations at the open pit deepened downwards to 1070mASL elevations, the initial tension cracks indicating a stability problem were observed in August 2009, at elevations between 1130 and 1155mASL of the southwest permanent slope (Fig. 2a). The total length of these cracks observed as two separate segments was measured to be about 125 m (MBEG, 2010). Although the mine planning initially envisaged the lowering of the groundwater level (GWL) down to the top of the coal, the observation of groundwater inflow from the SW slope and the open pit floor was considered as an indication that the expected result from the dewatering works could not be achieved.

3.2. The First Landslide

The crown of the mass moving on 8 January 2010 coincides with the line where the first cracks noticed in August 2009. It was determined that the groundwater intake into the mine came from new cracks that had formed on the SW slope adjacent to the Hurman River. In the site inspection reports, the main and small scarps were recorded, which were continuous on the benches and adjacent slopes (Fig. 2b). Cumulative vectoral displacements exceeding meters in the upper elevations and significant heaves in pit bottom have reported (Fig. 3). As can be clearly seen from the graph shown in Fig. 3, in some localities the weekly displacement reaches 1 meter, which indicates that mass movement continued during this period. Horizontal displacements in cracks have also measured at the beginning of January 2010. Mass movement has affected an approximately 13 hectares and the amount of displaced material estimated about 4.5-5.0 million cubic meters. Karpuz et al. (2010) suggested mitigation measures such as inside fill at toe area and excavation at top of SW slope to be taken to ensure stability on the Hurman River side in the oncoming phase of the project and emphasized the importance of the geological structure. By taking new findings into consideration, the slope stability calculations were revised based on the required measures.

After detailed site specific reports (MBEG, 2010, Karpuz et al., 2010), it was concluded that the continuation of mining activities without taking extra precautions would also pose a risk to the Belt Distribution Center and the Conveyor Belt Line. Therefore, the mechanism of landslide has been evaluated by analyzing 2D geotechnical models parallel to the direction of movement. After site inspection report, it was suggested that surface water leaking from existing cracks in the slope due to excessive rainfall during the autumn and winter periods reached the clay bands in and at the bottom of the coal, causing an increase in pore water pressure, and the first slide occurred throughout on swelled and lubricated clay lenses.

3.3. The February 2011 Landslides

3.3.1. Landslide I at Southwest Wall

In the monthly activity reports of November and December 2010, although no generally adverse situation was recorded in the fissures on the permanent western slope, small movements locally were mentioned. However, the January 2011 activity report stated the monitoring the cracks detected after the first landslide in 2010 on the permanent western slopes continued and movements were observed in some cracks (Park Teknik, 2011a). It was also noted there were swells of up to 1-1.5 m at the bottom of the open pit, and gas outflows and pressurized water discharge were observed through some cracks. This shows that the movement in the slope accelerated again in the second half of January. On February 6, 2011, at 03:30 am, the permanent SW slope was failed (Fig. 4a). Sliding mass having approximately length of 800 m has moved at southwestern slope system and displaced about 150 m in the direction of the excavation pit. The approximate volume of the mass displaced in this landslide was estimated as 20 million cubic meters.

In the crown area (MS and MSL in Fig. 4) of the Landslide I, a linear line-shaped rupture was observed in general and collapse zone (CZ) has occurred in front of the main scarp. The bedding plane orientations and mass integrity in the sliding block (SB) have been preserved, and there have been no significant changes in the geometry of the benches, slopes and mine transportation routes (Fig. 4b). Therefore, displaced block with original geological structures has been intended as an indication the shear surface can be planar (Fig. 4b). A collapse zone (CZ) with a width exceeding 100 meters has occurred between the main scarp of the landslide and the sliding block. In this zone, backwards toppling from the sliding block is evident (Fig. 4c) and large separation between the sliding block and the main rupture surface indicates also that the sliding surface can be planar. It was observed that the moving block caused a significant heave in the area of the toe and dragged the coal underneath it with the inner filling and leaned drifting materials locally against the opposite NE wall of open pit (Fig. 4d).

The locations of the monitoring points previously installed on the southwest slope and the measured cumulative vertical displacements between 1 December 2010 and 3 February 2011 prior to Landslide I are shown in Fig. 5. In this figure, the monitoring points shown on the map and the corresponding vertical displacements measured at these points on the graph are indicated with the same colours. The locations of the KN-4 and KN-6 monitoring points in this figure coincide with the main scarp of Landslide I. As seen from the graph, at these points, the sudden increase in vertical displacements on two different dates is evident (1 and 2 in Fig. 5). The first of these dates indicates reactivation of First Landslide, while the second indicates the acceleration just before Landslide I. Monitoring point, KN-13 is located on the horizontally displaced block (SB, in Fig. 4). The fact that no significant vertical displacement has been detected at this point can be considered to mean that the shear surface that limits the moving block from the base may be close to a horizontal position. In other words, the sliding surfaces of the First Landslide and Landslide I nay be the same surface, and stability problem in the southwest wall has been going on for a long time and accelerated about 10 days before the Landslide I.

3.3.2. Landslide II at Northeast Wall

Four days after the Landslide I, on February 10, 2011 at around 10:15 a.m. a second landslide (Landslide II) occurred on the northeast wall of the open pit. Unlike the first landslide, this movement was in the form of flow. The main scarp of Landslide II is semicircle with a radius of about 550 m, and its length in the direction of movement is about 1100 m. The estimated volume of the mass displaced by this landslide is 40–50 million m³ (Park Teknik, 2011b). After this movement, the displaced material overlapped the toe area of the Landslide I occurred before on the opposite slope and filled the open pit with 50 m thick material (Fig. 6 and Fig. 7).

The second movement happened in a very short time, collapsed in as little as 10 seconds, and was completed in about 5 minutes (Park Teknik 2011b). After this landslide, it was determined that a total of 10 people, including 2 expert engineers in the position of unit supervisor, were trapped under the landslide material and search and rescue operations were started. During the rescue efforts, the body of one worker was found and the remaining bodies of 9 people have not yet been found. Since then, lignite coal production activities have been suspended. Following the occurrence of the second landslide, the workplace was closed by the Ministry of Labour for all field activities (except drainage, rescue and stockyard activities) were halted. Considering the safe zone, dewatering activities were continued without entering possible danger zones. Drainage wells, which were deactivated due to lack of energy during landslides, were commissioned gradually. As of March 2011, a total of 297 drainage wells were regularly commissioned. In addition, the new dewatering well drilling activity continued with 2 drilling machines.

During site investigation, we observed floating blocks of the top cover (BLC in Fig. 7) rich in carbonate content on the wet blue clay matrix. Most of these blocks inclined backwards and are in the form of perfect curvilinear sequential lines, and partially embedded in the supersaturated blue clay. The displaced blue clay mass in the wet state is observed as an accumulation zone (BCM) of Landslide II. Another zone of mixed materials (MM&G) dragged forward and backward is found between two landslide masses. All these findings indicate that the sliding surface of the Landslide II should has a rotational component.

Considering the undulated structure shown on landslide material and spreading area of moving material (Fig. 7), a doubt has emerged that Landslide II may have been caused by an earthquake-induced liquefaction. For this reason, earthquake records up to 2 months ago were examined and their impacts on the mine site were discussed and evaluated. Earthquake records close to the mine site were compared temporally with geotechnical monitoring data, but no significant relationship was found.

3.4. Sinkhole Type Collapses

Karstic bedrock and the weak granular cover on it are another important boundary condition for dewatering work. Contacts between similar geological formations provide a balance between granular aquifers fed by surface water and sub-confined aquifers at the bottom and are susceptible to change in hydrogeological conditions. Excessive groundwater withdrawal from one of these aquifers causes interaction throughout their contacts. Where there is a similar geological relationship between two aquifers, if excess groundwater is discharged from one of these aquifers, the other tends to replenish the missing water. In the region of interaction between the two aquifers, a negative change occurs in the hydraulic gradient and thus in the stress state. As a result, a hydromechanically unstable situation can arise in the natural environment.

Following the two consecutive catastrophic landslides mentioned above (Landslide I and Landslide II), it was assessed that the groundwater level of the upper granular aquifer could not be lowered to the expected depth by ongoing dewatering works as the lower confined karst aquifer would feed the upper one. It has been argued that the pressure of the sub-aquifer should also be reduced in order to eliminate this negative situation (CU MMF, 2014, Ekmekçi and Yazıcıgil, 2016). After that, a second dewatering program targeting mainly the karst aquifer was planned and scheduled in February 2015. Within the scope of this dewatering project, additional dewatering wells in two rows almost parallel to the Hurman River were designed (Fig. 2). Some of these wells (coded with K in Fig. 8) were drilled and groundwater pumping from the karst aquifer has initiated in early February 2015. The total capacity of the pumps installed in the karst wells commissioned as of 12 February 2015 is 1447.2 m³/h. Between February 2015 and June 2015, a total of about 6.54 million cubic meters of water was discharged. The total amount of pumping water per month in May 2015 is 1.68 million cubic meters. This is 2.64 times more than the amount pumped in the same period of previous year (Fig. 9).

In the period from February to July 2015, 4 sinkholes (SH1, SH2, SH3, SH4) with different sizes and depths have formed consecutively at different time intervals in an area close to the edge of the basin, at distances 200–350 m from dewatering wells under operation (Fig. 8). The number of this circular shaped collapses is 7 by the end of 2015. The depths of them were different, but gradually increased by time in SH3 (Fig. 9).

It is indisputable that the failures such as caprock collapse type of sinkholes and surface cracks were occurred due to excessive groundwater withdrawal from the karstic aquifer (Fig. 11). Between 3 June 2015 and 16 June 2015, a sudden drop in the dynamic level of the K3 well closest to the sinkhole site was considered as a significant anomaly that may be associated with sinkhole formations (Fig. 11b). However, the efficiency of the K3 well is very low and the pumping was stopped at the end of July 2015. In addition, the piezometric level of the karst aquifer and the groundwater level of the upper aquifer were compared. According to the monthly dewatering activity reports corresponding this time, the stated groundwater levels are almost the same around the karst wells during this period. This is an important indicator of the interaction between two aquifers in this part of the Collolar coal field.

4.1. Geological Setting

The formation of the AEL basin is associated with erosional and depositional processes. The fact that the basin edges are irregular and do not contain any faulting data strengthens this view. Although Palaeozoic and Mesozoic aged sedimentary and metamorphic rocks locate on the basin edges, it is seen that carbonated rocks such as marble, limestone and calcschist surround the basin to a very large extent. Tertiary aged terrestrial and lacustrine deposits overlain unconformably the base rocks that also form the basin floor. Late Quaternary aged stream sediments and recent alluviums cover most of the surface of basin (Fig. 11). Extremely large karst cavities and related features are frequent in limestone and marble (Fig. 12).

The upper parts of the Plio-Quaternary basin fill overlaying the base rocks are evaluated with the help of our own detail observations at open pit walls, and the deeper layers under landslide masses were correlated using deep drilling data carried out for various purposes, and a generalized stratigraphic cross-section of the basin is obtained. This sequence rich in organic materials and consisting of clay, silty clay, coal and gyttja (a lacustrine mud containing abundant organic material) overlain unconformably the base rock at the southwestern side of the open pit. As it shown geological cross-section in Fig. 11, this formation starts with green clay (bottom clay) and ends with blue clay on top. Between these clay layers, a lignite horizon with varying thicknesses up to 100 m, gyttja with coal intercalations are continuous in a wide area. However, within the coal layer, there are high plastic clay and silt lenses with thicknesses varying from 0.5 to 1 meter. At open pit walls, inclinations of layers forming these lithological units are generally less than 5 degrees toward the north and northeast. The top blue clay is thinner in the southwest than in the northeast. In the southwest wall of pit, the black plastic clay/silt lenses within coal layer show low inclination (2°-5°) towards the pit floor and similar lenses also exist both inside the blue clay and gyttja units in the NE slope system. In the vicinity of southeast permanent slope, thin blue clay layer on gyttja at + 1150 ASL is continuous. However, it is eroded around the relocated riverbed and sinkhole area. In this part of Neogene basin, the bottom clay sitting on the base rock is either very thin or discontinuous. At the top, this sequence is covered with old and new alluviums or hillside remains. Although the quaternary formations are thicker on the northeast wall, a similar lithological sequence is also present in this wall.

4.2. Surface Water

Hurman River runs behind the SW slopes of open pit-mine, close to the top of the slopes which therefore constitutes a hydrologically important boundary condition for the southwest slopes. Its average annual flow rate is 8.5 m³/s and decreases to 1–3 m3/s in the dry season. During flooding, it reaches a flow rate of 45 m³/s (Karagüzel et al.,2018). After the last landslides in February 2011 riverbed was moved further southwest in order to protect the pit. The closest distance of the riverbed before February 2011 to permanent slopes is about 150 m. It was assessed that it would adversely affect the open pit, so the riverbed was displaced towards the edge of the basin (Fig. 8).

After the February landslides, the flow measurements made at stream gaging stations shown in Fig. 6, the flow rates in Hurman entry (SGS2), Kurudere entry (SGS3) and Hurman + Kurudere outlet (SGS1) were compared for the period of February and March 2011. The ratio between outlet (Q_SGS1) and inlet (Q_SGS2+Q_SGS3) discharges of surface water around open pit is generally higher than 1 during this period (top left of Fig. 6). Inferring that the discharges transferred from the mine pit to Hurman and the water pumped from the dewatering wells, it is considered that there was no significant water leakage from the riverbed around the open pit during this monitoring period.

4.3. Groundwater

The stratigraphic units mapped in and around open pit mine area are classified considering their physical and hydrogeological characteristics. In identifying hydrogeological systems with no significant differences regarding their hydrological characteristics, the gyttja units with coal interlayers which were called as “carbonate gyttja” in previous studies are together classified as gyttja in this study. Similarly, the coal units with thin gyttja interlayers that were previously identified as “coal bearing gyttja” are considered within “coal” layers (Table 1).

The hydraulic conductivities of base rock and upper layers in sinkhole area were obtained from pumping and seepage tests. Within the scope of this study 3 additional geotechnical boreholes have been drilled extending down to bedrock. During the pumping tests these 3 boreholes were also used as observation wells. There are many data obtained previous pumping tests, the results of which are also considered when evaluating hydrogeological systems (Table 2). There are significant differences in the average values of the hydraulic parameters in the sinkhole area.

4.4. Dewatering

For dewatering the upper layers of Plio-Quaternary sequence, perforated S-coded wells close together along slopes and benches were commissioned synchronically with excavation (Fig. 2). From zero measurements in these dewatering wells, it was determined that the GWL was shallow and its depth ranged from 0 to 10.17 meters. In fact, in more than 80% of 157 wells, the depths of GW were found to be less than 5 m. However, the hydraulic conductivities of hydrogeological units are low due to the fact that they contain a significant amount of clay and silt. Therefore, the radius of cones of drawdown are very narrow in these wells. Considering the total operating time in the period between 2007 and 2012, it was determined that these wells were active at only 10–50% of their capacity. On average, 53 percent of the pumps have remained active for 50 percent of the total time. Approximately 25% of pumps are active for less than 20% of the total operating time, while 8% are active for less than 10% of the same term. Such cases are usually caused by insufficient water flowing into the wells.

In order to check the success of the dewatering works, groundwater level (GWL) monitoring program was revised in August 2009 after the first cracks in the upper bench of the southwest slope. However, the true depth of the groundwater is not known exactly, since there are no observation wells on the pit floor. Nevertheless, using measurements from existing observation wells and assuming that the GWL is theoretically lowered to the bottom of the open pit. Then measured levels in May 2010 compared with wellhead elevations, top level of coal horizon and riverbed elevation of Hurman (Fig. 13). It is clear that the measured drawdowns in observation wells are very different and are mostly higher in those with higher wellhead elevations. It is minimum in the S 445 well (located behind the NE wall). However, the GWL measured in all observation wells is above the top of coal horizon. In other words, the groundwater table have not been lowered down to the coal horizon, which was foreseen in slope stability calculations. In brief, the yield of gyttja wells is low and the dewatering program targeting the upper aquifer has not realized in May 2010. The same comparison was made for previous months before landslide disaster in February 2011 and it was concluded that the expected efficiency from the dewatering program could not be achieved.

4.5. Seismicity

Before the landslides, according to the Kandilli Observatory's Regional Earthquake and Tsunami Monitoring Centre, two earthquakes occurred between February 4 and February 11, 2011 at distances of 50 and 141 km far from the open pit mine. Their magnitudes are 2.9 and 3.6 respectively. Maximum horizontal ground accelerations (Cornell et al., 1979) that these earthquakes would cause in the study area were calculated and determined to be below 0.002g. It was evaluated that these earthquakes would not be destructive and would not have triggered landslides at the mine site.

4.6. Geotechnical Model

From numerous drilling logs and field observations, it has been determined that the relationship between the layers of the Neogene sequence is simple and continuous in Collolar coalfield. The geological models of the pit slopes and the sinkhole area were prepared by correlating the borehole logs. Field observations and new core drillings within the scope of the study are also considered. Geotechnical properties of model layers were defined based on the results of the laboratory and in-situ tests.

Within the scope of the study, core samples were taken from the boreholes numbered B-JS 34 and B-JS 05 which were cutting all geotechnical units in the zones of southwest and northeast slopes respectively (Fig. 2). These samples were subjected to sieve analysis and consistency limit tests the results of which are given in Table 3. Mineralogical analyses through X-ray Diffraction (XRD) were conducted on 13 samples. In general, XRD analysis show good compliance with those of consistency limit tests.

UD samples taken from the geotechnical boreholes around sinkhole area were subjected consolidated undrained (CU) direct shear tests in soil mechanics laboratory of civil engineering department of ITU. However, for the selection of geotechnical model parameters, peak and residual shear strength parameters obtained from previous surveys were also considered (Table 4). Some layers of geotechnical model are not present in the sinkhole area.

5.1. Selection of Geotechnical Parameters

The assessments suggest that in open-pit area, residual internal friction angles (φ_r) of bottom clay and black clay layers within coal could be taken as 12^° and 9° respectively. The geotechnical parameters representing other geological units were selected from the data provided from previous investigations and this study. All mentioned failures in open pit walls are sequentially analysed based on geotechnical properties given in Table 5, and a conceptual model of sinkhole area was analysed considering the karst dewatering activity initiated behind the southwest permanent slope of the open pit mine.

5.2. Instabilities of SW Wall of Open Pit

In order to reveal the mechanism underlying the instability problem in the southwest slope system, FEM analyses repeated for each case following.

Phase 1: August 2009 (the First Cracks)

Phase 2: January 2010 (the First Landslide)

Phase 3: February 6, 2011 (Landslide I)

The most critical 2D models of this slope system were transferred to the Rs2 version: 11.020 software (Rocscience, 2023). Existing slope geometries and groundwater levels were defined separately at each phase and geological models were constructed by correlating borehole logs on or near cross-section lines. Except base rock, the Mohr-Coulomb failure criterion was used for model layers. The base rock was introduced to model according to Hoek-Brown failure criterion embedded to this software. For all phase, the level of the riverbed of Hurman and the elevations of the open pit floors were considered as constant hydraulic levels. The external boundary of models was fixed at bottom and free at both sides. Six noded triangular uniform mesh selected for discretization of FEM models. The ratio of field stresses (σh/σv) has considered as 0.40 for all cases.

Using the Shear Strength Reduction (SSR) method (Hammah et al., 2005) the stability of soil or rock slopes can be determined by finding the critical strength reduction factor (SRF). The Shear Strength Reduction option in RS2 is used for cases implemented in this study. It allows us to automatically perform a finite element slope stability analysis, and compute a critical SRF for the model. The critical strength reduction factor is equivalent to the "safety factor" of the slope (Rocscience, 2023).

The 2D Geological model, and some results from the analysis of the Phase 1, are shown in Fig. 14. The critical strength reduction factor (SRF) was obtained as 0.93 for this phase. It has been seen that the maximum shear strain develops along two planes close to the horizontal, which correspond to the contact between the base rock (BR) and the bottom clay (BTC), and the main shear zone (SZ) appears to have curved backwards and lengthened to the area where the first surface cracks (FC) were formed, which were detected in August 2009 (Fig. 2a). Despite the fact that the black clay layer within coal (BLC) has a relatively low shear strain, it can also be considered to have an effect on the stability of the SE slope. However, for critical SRF, the maximum total displacement along the shear zone is much higher than what was observed in August 2009. But it is approximately 1 m for SSR value of 0.75.

Similar results were found out from repeated analysis for the Phase 2, when displacements in August 2009 were not considered (Fig. 15). The slip zone almost coincides with the one obtained for previous phase, and while the critical SRF value decreased to 0.73, the total deformations increased and the movement of the block on the black clay towards the bottom of the pit became evident. In addition, due to the deepening of the excavation pit, a heave up to approx. 1m high occurs at the pit floor. The FEM analysis indicates that translational motion of a block on black clay and bottom clay has continued since August 2009. This confirms our field observations, which we have already discussed under previous headings.

Figure 16 shows the results of third phase of FEM analysis concerning the stability condition in the southwest slope just before the Landslide I. Decrease in critical SRF value obtained for this phase to 0.70 indicates that the SW slope system has been progressively deteriorating since August 2009. The vertical displacement corresponding to the critical SRF value reaches 30 m, while the horizontal displacement reaches 36 m. The subsidence zone in this figure and the block displaced horizontally in front of this zone into the open pit completely coincide with our own field observations (Fig. 4b and 4c). As mentioned above, despite the mitigations such as earth removal work at the top area of this slope and in pit filling to support the toe zone, the mass movement starting in August 2009 could not be prevented. Mine activity reports indicate that coal production continued on the bottom of the open pit on this date and excavations were carried out that reached even below the lowest level. Also, before the landslide, the Coal Panel F (Fig. 2b and Fig. 18) in the toe area of the NE wall supporting each opposite slope was entirely excavated away. In brief, the analysis of the 2D FEM models corresponding to SW slope system suggest that that the translational movement of the same block, which began in August 2009 on the bottom clay and black clay, moved until February 2011, while the displacement rate changed from time to time.

5.3. Instabilities of NE Wall of Open Pit

The most critical and representative geological models of NE wall were set up using borehole data for two different cases (Fig. 17a and 18a). However, since there is no drilling in this region extending to the bedrock, the thickness of the bottom clay is estimated to be 50m, and the boundary of the bedrock and the bottom clay is assumed to be horizontal. The first case corresponds to the 2D slope geometry in January 2010, while the second in January 2011.

As can be seen from Fig. 17, the 120m wide F coal panel that ensures the stability of the slope in the toe area of the northeast wall has not yet been excavated in January 2010 and the pit floor elevation was approximately 1050mASL. From the results of the 2D FEM analysis of this phase, it becomes clear that the slope of the NE is in a critical equilibrium state (SRF = 1), and there is no displacement in this wall. But a potential shear zone appears along a semi-circular surface that also passes under the coal panel F.

The results of FEM analysis, which was renewed considering the condition of the same wall in January 2011, are shown in Fig. 18. The location of the tension crack (TC) observed before Landslide II (MTA, 2017) at the top bench crest is marked on the geological model. From the excavation profile in this figure, it is evident that the coal panel F in toe zone was completely excavated on this date. For the critical SRF value of 0.63, the shear zone and total displacements are shown in Fig. 18b and Fig. 18c respectively. This solution yields similar components that demonstrate a rotational motion and backward tilting mechanism that we pointed in Fig. 7. In addition, Fig. 18c refers to a potential tension zone (TZ) of about 500 m, which corresponds to a zone of successive failures extending backwards behind the upper bench top after Landslide II. The results from the FEM analyses of the northeast wall coincide significantly with our field observations in the area where Landslide II occurred. Without reducing the overall angle of the northwest wall and reducing the GWL to the level envisaged in mine planning, it shows that the excavation of the F panel adversely affected the stability of this wall.

5.3. Karst Drainage Related Collapses

Within the scope of this study, additional boreholes were drilled to determine the geotechnical characteristics of the model layers in the zone where the mentioned sinkholes were formed. Disturbed (SPT) and undisturbed (UD) samples were taken from different depths of these boreholes, and they were subjected to sieve analysis, consistency limit and consolidated and drained (CD) shear tests. Some of the test results are given in Table 6. It was determined that the alluvium composed of coarse gravel with small amount of sand, and sand content did not exceed 25%. The thin layer of cohesive top soil covering the alluvium in the sinkhole area has medium plasticity (MI), while samples from gyttja have high plasticity (CH, MH). The natural water content of the UD-3 sample taken from the depth of 37.50–38.00 m of the ITU-3 borehole is higher than the liquid limit of this sample.

As can be seen from Table 6, the porosity of the gyttja is usually over 60%, and peak shear strength parameters obtained for gyttja varies in a wide range. If shear strength parameters compared to the MBEG (2012) data presented in Table 4, it is understood that the same layers of Plio-Quaternary sediments are weaker at the edge of the basin.

The main boundary conditions of a conceptual model based on field observations and drillings in sinkhole area are shown in Fig. 19. Theoretically, it was hypothesized that excessive discharge from new karst wells could cause an interaction between two aquifers that meet at the edge of the basin. It was also considered in this model that there may have been shallow karst cavities (KC) in the bedrock in the area where the sinkholes formed.

The conceptual model was checked in two different ways. In the first, groundwater was pumped from K-coded karst wells whose locations are marked in Fig. 8, and then groundwater levels were measured in the boreholes close to the sinkholes. The measured levels were also correlated considering the levels measured at research pit and riverbed elevations. This comparison shows that the groundwater level (GWL) decreases towards the ITU-1 borehole at the edge of basin and the flow line develops in the same direction. It was found out that during the withdrawal of groundwater from the karst wells, the cone of drawdown would be in the area where the sinkholes are clustered, therefore this comparison supports the conceptual model. 2D FEM analysis was used as a second way to check the validity of the conceptual model. For this purpose, the conceptual model was digitized and analyzed using data from geotechnical boreholes and pumping experiments. Geotechnical parameters and hydraulic conductivities of model layer were selected from the Table 2, Table 5 and Table 6. Hydraulic head of confined aquifer and the Hurman River are introduced separately. In the numerical model, the water level of the Hurman River feeding the upper alluvium aquifer is kept constant. At the last stage of modeling, karst dewatering process was introduced to model.

The average filtration rate of karst wells which were under operation calculated according to the equation below.

$$V=Q/\left({F}_{k}{H}_{f}\pi \text{r}\right)$$

Where;

$Q$ : Monthly total discharge from karst wells (865670 m3/month) in March 2015

${F}_{k}$ : Filter coefficient = 0.1

${H}_{f}$ : Total length of filtered zones in karst wells under operation at the same period = 104m

$\text{r}$ : Radius of karst wells = 0.5m

In this model, a karstic cavity which is not in regular geometry and full of groundwater randomly generated in shallow depth of base rock (Fig. 20a). After performing the FEM analysis, no significant displacement occurred in the area where the sinkholes formed. However, considering the karst structure of the bedrock, only the initial hydraulic conductivity of the karst aquifer was increased, without changing the other model parameters, until the vertical displacement observed in the cavities was reached. Because of this calibration; the hydraulic conductivity of karst aquifer has been increased up to 2.32E-2 m/s.

After starting groundwater discharge from the sub-aquifer, it was determined that both the hydraulic gradient and the velocity of the groundwater increased, while the pore pressure decreased below the area where the sinkholes formed (Fig. 20c and 20d). FEM analysis also shows that groundwater flow lines are controlled by contact between alluvial and base rock, while vertical displacements up to 3.5 m occur in the area where the sinkholes are located (Fig. 20e).

The AEL coal basin, which has an indispensable place in thermal energy production in Türkiye, has unique difficult and challenging conditions for mining activities. With more than 20% of the country's proven lignite reserves, this coal basin has long been a focus on vital importance of dewatering and slope stability issues. Despite this, the consecutive landslides in February 2011 in open pit walls in the Collolar sector and serious property and life losses occurred.

In this study, the stability problems in the Collolar Open Pit are discussed and their interactions with each other are examined. The instabilities encountered in and around the open pit slopes were evaluated in a holistic approach in chronological order, considering observation, measurement, experiment and monitoring data and mining activities such as dewatering and excavation. Two-dimensional numerical models of the sinkhole area, and the southwestern and northeastern walls of the open pit corresponding to different time periods were set up and analyzed separately. Basically, the following main conclusions have been drawn from the 2D FEM analyses and assessments based on the major characteristics of the geological environment.

In the SW slope of the open pit, the first crack detected in August 2009, the first landslide in January 2010 and Landslide I in February 2011, which caused loss of life and property, have the same mechanism of failure. 2D FEM analyses of SW wall of the open pit show that Landslide I is an end of movement of the same block sliding translationally on bottom clay and black clay.

The results of the FEM analysis of the Landslide II in NE wall coincide significantly with our field observations and borehole data after this landslide. FEM analysis gives the typical components of a rotational movement. It also indicates a tension zone of 500 m behind of top bench crest of this slope where successively slices of landslide could clearly observed, and adverse effect of the excavation of the F-panel supporting this wall. The rapid expansion of the collapse zone can be explained by supersaturation of weak geological layers above the lignite horizon.

It has been determined that due to excessive groundwater pumping, hydraulic equilibrium may be disturbed and an interaction may develop along the contact between the upper hydrogeological system and the karstic sub-aquifer at the edge of the basin. In such case, both the hydraulic gradient and the velocity of the groundwater increase, while the pore pressure (bouncy support) decrease simultaneously. FEM analysis showed that sinkholes formed after excessive groundwater withdrawal from the sub aquifer in the area under study may be associated with the failure of the roofs of one or more shallow karst cavities within the base rock and soil piping in the weak cover of the Neogene sequence.

From the back analyses, it was concluded that the high plasticity clay layers within geological formation and the hydraulic interaction of two aquifers with different characteristics are two another key factor controlling mining activities at the edge of the AEL basin.

Author Contribution

Mahmutoğlu wrote the manuscript text and draw the most of figures and some tables. He also made the numerical analysis.Karaguzel worked on the field geological survey and hydrogeology and studied on dewatering data.

Acknowledgements

We thank to the Electricity Production Co. Inc. of Türkiye (EÜAŞ) for financial supports of this project and providing valuable data. We would also like to thank the Foundation of the Faculty of Mines for giving us the opportunity to conduct this project work, and Dr. Caner Zanbak for his valuable comments.

Akbulut I, Çam I, Aksoy T, Cağlan D, Ölmez T (2013) Slope instability in open pits and an example of a retrospective analysis: Afşin-Elbistan-Kislakoy Open Pit Coal Deposit. Bulletin of Mineral Research and Exploration Institute of Turkey147: 115-126.
Akçar N, Yavuz V, Ochs SI, Fredin O, Nyffenegger F, Korkut M, Margreth A, Gylland AS, Schlunegger F (2019) Liquefaction of freshwater carbonates led to the February 10, 2011, landslide at the Cöllolar coalfield, eastern Turkey. Geomorphology 347:106859.
Besbelli B, Karaca K, Gökmenoğlu O (2009) Geology, reserve and hydrogeology of HB and HD sectors of Afşin-Elbistan coal basin. Energy Department of General Directorate of Mineral Research and Exploration (MTA) of Turkey. Archive No: 11238 (Unpublished Report in Turkish).
CU MMF (2015). Integrated dewatering project for Collolar Coal Basin, Çukurova University, Faculty of Engineering and Architecture, Adana (Unpublished Report in Turkish).
Cornell CA, Banon H, Shakal AF (1979) Seismic motion and responce prediction alternatives, Earthquake Engineering and Structural Dynamics V.7, No:4, 295-315.
Ekmekçi M, Yazıcıgil H (2016) Progressive sinkhole occurrence induced by dewatering activities in a large lignite mine (SE Turkey). Proceedings IMWA 2016, Freiberg/Germany, Drebenstedt, Carsten, Paul, Michael (eds.), Mining Meets Water – Conflicts and Solutions, pp 486-493.
Gökmenoğlu O, Aslan M (2013) Hydrogeological investigation project of Kahramanmaraş Afşin-Elbistan coal basin, Dewatering report for east side of Hurman River. General Directorate of Mineral Research and Exploration (MTA) Energy Department, Achieve No: 1346, Ankara (Unpublished report in Turkish).
Gutierrez F, Guerrero J, Lucha P (2008) A genetic classification of sinkholes illustrated from evaporite paleokarst exposures in Spain. Environ Geol 53:993–1006.
Gürsoy E, Özcan K, Yücel AR (1981) Coal deposit geology report of Kahramanmaraş Elbistan D1 Sector. General Directorate of Mineral Research and Exploration Report No: 7054, (unpublished), Ankara.
Hammah RE, Yacoub TE, Corkum BC, Curran JH (2005) The shear strength reduction method for the Generalized Hoek-Brown Criterion, Alaska Rocks 2005, The 40th U.S. Symposium on Rock Mechanics (USRMS): Rock Mechanics for Energy, Mineral and Infrastructure Development in the Northern Regions, Alaska.
Karagüzel R, Akyüz SH, Mahmutoğlu Y, İşseven T, Bozkurtoğlu E (2017) Geological research for sinkhole occurrence at Afşin-Elbistan Coal Basin, ITU Mining Faculty Foundation, pp 217 and appendixes, Istanbul (Unpublished Report in Turkish).
Karagüzel R, Akyüz HS, Mahmutoğlu Y, İşseven T, Tunçdemir H, Bozkurtoğlu E, Şans G, Erdoğan M, Doğan A, Özer Ü (2018) Investigation of the mechanism of February 2011 Landslides in Çöllolar Mine Site in Afşin Elbistan Coal Basin, ITU Mining Faculty Foundation, (Unpublished Report in Turkish) Istanbul.
Karpuz C, Tutluoğlu L, Öge, İF (2010) Elbistan-Çöllolar Open Pit operation landslide analysis and solution suggestions, (Unpublished Report in Turkish) Ankara.
Kavvadas M, Roumpos C, Schilizzi P (2018) Investigation of the stability of deep excavation slopes in continuous surface lignite mines. Proceedings, 14th International Symposium of Continuous Surface Mining (ISCSM) Thessaloniki Greece.
Lashgari M, Ozturk CA (2022) Slope failure and stability investigations for an open pit copper mine in Turkey. Environmental Earth Sciences 81:5.
Leonardos M (2004) Methods and procedures of monitoring, assessment and improvement of the excavated slopes stability in the deep Greek lignite exploitations, PhD Dissertation, National Technical University of Athens.
Mahmutoğlu Y, Karagüzel R, Sans G, Erdoğan M, Bozkurtoğlu E, Akyüz HS, İsseven T, Taştekin C, Ata E (2017) The effect of karst dewatering on sinkhole occurrence, A case study from Afşin-Elbistan Coal Basin, National Symposium of Engineering Geology and Geotechnics, proceeding book, In: Çan T and Taga H (Eds) 164-171, Adana.
Mahmutoğlu Y, Karagüzel R, Sans G, Topcuoğlu M (2019) Sinkholes induced by dewatering in an open-pit mine: Case study from a coal basin in Eastern Turkey. Rock Dynamics Summit, Aydan O, Ito T, Seiki T, Kamemura K and Iwata N (Eds). 597-603. Okinawa, Japan.
MBEG (2008) Collolar Open Cast Mine Afsin Elbistan Coal Basin, Phase II: Detail Planning Çöllolar Open Cast Mine Subproject 2: Hydro-geological model and dewatering concept, unpublished report.
MBEG (2010) Çöllolar cast mine, Afşin-Elbistan Coal Basin, Phase IV/2010: Update of Detail Planning Collolar Open Cast Mine: Step1: Preliminary report.
MBEG (2012) Consultancy and engineering services for mine operation in open cast mine (Afşin-Elbistan Coal Basin), Phase V: Mine plan for the restart of coal mining & status report.
MTA (2017) 5-year operation plan of Kahramanmaraş Afşin-Elbistan coal basin sector B (Collolar coalfield) based on general slope design. General Directorate of Mineral Research and Exploration of Turkey, Project No: 2016-34-41 (Unpublished report in Turkish).
Newton, JG (1984) Sinkhole resulting from groundwater withdrawals in carbonate terranes -an overview. Geological Society of America, Reviews in Engineering Geology, volume VI. 195–202.
Özbay A, Cabalar, AF (2015) FEM and LEM stability analyses of the fatal landslides at Collolar open-cast lignite mine in Elbistan, Turkey. Landslides,12:155–163.
Park Teknik (2011 a) January 2011 Monthly Activity Report (in Turkish).
Park Teknik (2011 b) February 2011 Monthly Activity Report (in Turkish).
Rocscience (2023) Rs2 v11.020, 2D Geotechnical finite element analysis, Registered to Istanbul Technical University, Rocscience Inc., Toronto, Ontario, Canada.
Schloemer, W (1972) Hydrogeologische Voruntersuchungen im Bereich des zukünftigen Braunkohlen-Tagebaues bei Elbistan. Geologisches Jahrbuch, B.4, S.53-136, Hannover.
Sevil, J, Gutierrez F (2023) Morphometry and evolution of sinkholes on the western shore of the Dead Sea. Implications for susceptibility assessment. Geomorphology 434 108732.
Tharp, TM (1999). Mechanics of upward propagation of cover-collapse sinkholes. Engineering Geology 52, 23–33.
Tutluoğlu L, Öge IF, Karpuz C (2011) Two and three-dimensional analysis of a slope failure in a lignite mine. Computers & Geosciences 37(2):232–240.
Tzampoglou P, Loupasakis C (2023) Hydrogeological hazards in open pit coal mines–investigating triggering mechanisms by Validating the European Ground Motion Service Product with Ground Truth Data. Water 15, 1474.
Wolkersdorfer C, Thiem G (1999) Ground water withdrawal and land subsidence in north eastern Saxony (Germany). Mine Water Environ 18, 81–92.
Yechieli Y, Abelson M, Baer G (2016) Sinkhole formation and subsidence along the Dead Sea coast Israel. Hydrogeology Journal 24 (3), 601–612.
Yörükoğlu M (1991). Afsin-Elbistan mining project and mining activities at AELI Establishment of TKİ. Turk Min J. 30(3):13–29.
Zevgolis IE, Deliveris AV, Koukouzas NC (2019) Slope failure incidents and other stability concerns in surface lignite mines in Greece. Journal of Sustainable Mining 18:182–19.

Table 1. Hydrogeological systems and their characteristics in open pit area.

Formation	Lithology	Hydrogeological System	Types of Aquifers
Alluvium	Recent and old Alluviums	Coarse (Poorly cemented)	Unconfined
Alluvium	Recent and old Alluviums	Coarse (Poorly cemented)	Unconfined
Fine Clastic with Coal	Top Clay	Cohesive	Aquitard
	Marl	Consolidated	Aquitard
	Gyttja	Coarse/Cohesive	Poor Aquifer/Aquiclude Unconfined/Semi Confined
	Coal	Fissured/Jointed
	Lower Gyttja	Coarse/Cohesive
	Bottom Clay	Cohesive	Aquitard
Coarse Sediments (Eocene)	Intercalation of Sandstone, Claystone and Coarse Sand	Coarse Poorly Cemented Layered Jointed	Semi High Confined
Base Rocks	Limestone Marble Schist	Karstic and Fractured	Extensive-High Confined

Table 2. The hydraulic conductivities of model layers obtained from pumping, lugeon and seepage tests around the sinkholes area.

Model Layer	k (m/s)
	MBEG, 2012		Sinkhole Area (This Study)
	Arithmetic Mean	Geometric Mean	Arithmetic Mean	Geometric Mean
Alluvium	3.33E-05		3.09E-4	2.92E-4
Top Clay	4.31E-06	4.27E-6	-	-
Gyttja	8,80E-07	8.52E-7	3E-7	3E-7
Coal	1.15E-07	1.73E-7	1.58E-4	1.58E-4
Bottom Clay	1.15E-06	1.05E-6	-	-
Base Rock (Limestone)	1.00E-3		2.98E-4	2.97E-4

Table 3. The results of consistency limit tests, sieve and XRD analysis of core samples taken from geotechnical boreholes B-JS5 (Northeast) and B-JS34 (Southwest). Q: Quartz, F: Feldspar, C: Calcite, I: Illite, Cl: Chlorite, S: Smectite.

BH No	Depth (m)	Sample No	Consistency Limits (%)			Sieve Analysis in Wet Condition		XRD Analysis Minerals by Wt. (%)						Model Layer
BH No	Depth (m)	Sample No	LL	PL	PI	+ #200	−#200	Q	F	C	I	Cl	S	Model Layer
B-JS 05	10.60-11.00	N1	58	25	33	11.99	88.01	20-25	-	45-50	-	<5	25-30	Loam
	19.40-19.70	N2	76	32	43	1.02	98.98	25-30	-	45-50	-	<5	20	Loam
	28.80-29.00	N3	59	28	31	10.22	89.78	15-20	<5	30-35	-	5-10	30-35	Blue Clay
	35.70-36.00	N4	104	44	60	5.24	94.76	20-25	-	25-30	-	10-15	35-40	Blue Clay
	42.70-42.90	N5	58	23	35	22.10	77.90	<5	-	90-95	<5	<5	-	Beige Gyttja
	77.80-78.00	N6	51	41	10	7.96	92.04	-	-	-	-	-	-	Grey Gyttja
	107.2-107.4	N7	113	29	84	11.63	88.37	-	-	-	-	-	-	Black Clay
	167.3-167.5	N8	58	29	29	2.70	97.30	20-25	<5	35-40	15-20	15-20	-	Bottom Clay
B-JS 34	19.00-19.30	N-6	84	41	43	4.17	95.83	20-25	-	25-30	-	-5-10	40-45	Blue Clay
	90.05-91.50	N-5	NA			NA	-	-	-	-	-	-	-	Coal
	112.7-112.9	N-4	61	34	27	11.67	88.33	60-65	<5	<5	5-10	15-20	5-10	Black Clay
	114.0-114.3	N-3	56	26	30	0.56	99.44	50-55	5-10	15-20	5-10	10-15	-	Cay
	128.2-128.4	N-2	42	23	19	2.58	97.42	25-30	5-10	15-20	20-25	20-25	-	Black Clay
	146.6-146.7	N-1	42	25	17	0.72	99.28	10-15	-	75-80	<5	<5	<5	Bottom Clay

Table 4. Comparison of peak and residual shear strength parameters of model layers obtained from consolidated undrained shear tests. Natural unit weight (γ_n), peak and residual cohesions (c_p and c_r) and peak and residual friction angles (Φ_p and Φ_r) are in kN/m³, kPa and degree (°) respectively.

Model Layers	References
	MBEG, 2012					MTA, 2017					ITU (This study)
	γ_n	c_p	c_r	Φ_p	Φ_r	γ_n	c_p	c_r	Φ_p	Φ_r	γ_n	c_p	c_r	Φ_p	Φ_r
Alluvium	-	-	-	-	-	-	-	-	-	-	18.0	0	0	35	-
Loam	17.3	31	22	27	19	18.3	36	21	26	21	18.1	20	-	31	-
Blue Clay	18.5	42	31	16	9	17.9	66	49	20	16	-	-	-	-	-
Grey Gyttja	13.9	25	15	32	30	15.0	51	43	29	28	14.5	5	-	35	-
Beige Gyttja	13.6	38	14	30	29	15.7	39	32	32	32	14.3	20	-	34	-
Coal	11.8	30	21	33	27	11.5	122	79	19	15	-	-	-	-	-
Black Clay	17.2	46	15	15	7	17.3	72	37	14	12	-	-	-	-	-
Bottom Clay	16.6	32	15	17	9	18.6	77	58	20	16	-	-	-	-	-

Table 5. Peak and residual shear strength parameters of model layers.

Model Layers	Geotechnical Properties
Model Layers	Φ_p (°)	c_p (kPa)	γn (kN/m³)	Φ_r (°)	c_r (kPa)
Inside Fill and Damp	27	5	15.0	20	1
Alluvium	35	0	18.0	27	0
Loam	28	29	17.8	20	22
Blue Clay	18	54	18.2	12	45
Black Clay	15	59	17.3	9	26
Gyttja	32	32	14.5	30	23
Coal	26	76	11.7	21	50
Bottom Clay	19	54	17.1	12	36

Table 6. Solid and natural unit weights (γ_s, γ_n), porosity (n) and peak shear strength parameters obtained from consolidated drained (CD) tests of UD samples taken from ITU boreholes drilled sinkhole area.

Units	γ_s(kN/m³)	n	γ_n(kN/m³)	c_p (kPa)	Φ_p (◦)
Top Soil	27.0	0.48	18.1	20	31
Beige Gyttja	23.6	0.61	14.3	20	34
Gyttja	24.5	0.59	15.8	60	33
Beige Gyttja	24.3	0.63	14.5	5	35
Beige Gyttja	-	-	15.0	5	35
Gyttja with Coal	25.1	0.71	13.6	35	32

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

The Key Factors Controlling the Mechanisms of Consecutive Failures in the Collolar Open Pit in the Afsin-Elbistan Lignite (AEL) Basin in Southeast Türkiye

Status:

Version 1

Abstract

Figures

1. INTRODUCTION

2. GEOGRAPHIC LOCATION

3. THE CONSECUTIVE INSTABILITIES IN THE COLLOLAR OPEN PIT SITE

3.1. The First Tension Cracks

3.2. The First Landslide

3.3. The February 2011 Landslides

3.3.1. Landslide I at Southwest Wall

3.3.2. Landslide II at Northeast Wall

3.4. Sinkhole Type Collapses

4. BOUNDARY CONDITIONS IN OPEN PIT SITE

4.1. Geological Setting

4.2. Surface Water

4.3. Groundwater

4.4. Dewatering

4.5. Seismicity

4.6. Geotechnical Model

5. NUMERICAL ANALYSES OF CONSECUTIVE FAILURES

5.1. Selection of Geotechnical Parameters

5.2. Instabilities of SW Wall of Open Pit

5.3. Instabilities of NE Wall of Open Pit

5.3. Karst Drainage Related Collapses

6. CONCLUSIONS

Declarations

Author Contribution

Acknowledgements

References

Tables

Additional Declarations

Status:

Version 1